Confusion in using BicubicSplineInterpolation

[amar singh]

I observed a strange thing in using the class. If the length of X-vector supplied to the operator() is >= Y-vector length ,it gives proper z-values at all the points. However, I get wrong z-values are obtained at some points if X-Vector length < Y_vector length, and those points seem to be related to the difference between the two vector lengths.

eg., In the following code,I assign z-value=1 at all the nodes in the 5X6 grid.All points should give an interpolated z-value of 1 from this function. But there are some points which have a random value .It works OK if I make it 6X5.

 

Am I using it wrongly, or could there possibly be some problem?

 

Regards,

Amar

 

Following is the code:

    {
      std::vector<double> x(5);
      x[0]=0;
      x[1]=1;
      x[2]=2;
      x[3]=3;
      x[4]=4;
      //x[5]=5;
      std::vector<double> y(6);
      y[0]=0;
      y[1]=1;
      y[2]=2;
      y[3]=3;
      y[4]=4;
      y[5]=5;
      Math::Matrix z(x.size(),y.size());
      int i,j;
      std::vector<MyExcelRange::StringVectorType> matrix;
      for(i=0;i<x.size();i++)
      {
        for(j=0;j<y.size();j++)
        {
          z[i][j]= 1;
        }
      }

      ofstream myfile("outfile2.txt");
      myfile << "**************************  before " << endl;
      for(i=0;i<x.size();i++)
          for(j=0;j<y.size();j++)
            myfile << "x = " << x[i] << " y = " << y[j] << " z = " << z[i][j] << endl;

      myfile << "**************************  from interpolation" << endl;
      for(i=0;i<x.size();i++)
          for(j=0;j<y.size();j++)
            myfile << "x = " << x[i] << " y = " << y[j] << " z = " << Math::BicubicSplineInterpolation<std::vector<double>::const_iterator,std::vector<double>::const_iterator,Math::Matrix> (x.begin(),x.end(),y.begin(),y.end(),z)(x[i],y[j]) << endl;
      
              }

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Following is the output:

**************************  before
x = 0 y = 0 z = 1
x = 0 y = 1 z = 1
x = 0 y = 2 z = 1
x = 0 y = 3 z = 1
x = 0 y = 4 z = 1
x = 0 y = 5 z = 1
x = 1 y = 0 z = 1
x = 1 y = 1 z = 1
x = 1 y = 2 z = 1
x = 1 y = 3 z = 1
x = 1 y = 4 z = 1
x = 1 y = 5 z = 1
x = 2 y = 0 z = 1
x = 2 y = 1 z = 1
x = 2 y = 2 z = 1
x = 2 y = 3 z = 1
x = 2 y = 4 z = 1
x = 2 y = 5 z = 1
x = 3 y = 0 z = 1
x = 3 y = 1 z = 1
x = 3 y = 2 z = 1
x = 3 y = 3 z = 1
x = 3 y = 4 z = 1
x = 3 y = 5 z = 1
x = 4 y = 0 z = 1
x = 4 y = 1 z = 1
x = 4 y = 2 z = 1
x = 4 y = 3 z = 1
x = 4 y = 4 z = 1
x = 4 y = 5 z = 1
"**************************  from interpolation

x = 0 y = 0 z = 1
x = 0 y = 1 z = 1
x = 0 y = 2 z = 1
x = 0 y = 3 z = 1
x = 0 y = 4 z = 1
x = 0 y = 5 z = -1.45682e+144
x = 1 y = 0 z = 1
x = 1 y = 1 z = 1
x = 1 y = 2 z = 1
x = 1 y = 3 z = 1
x = 1 y = 4 z = 1
x = 1 y = 5 z = -1.45682e+144
x = 2 y = 0 z = 1
x = 2 y = 1 z = 1
x = 2 y = 2 z = 1
x = 2 y = 3 z = 1
x = 2 y = 4 z = 1
x = 2 y = 5 z = -1.45682e+144
x = 3 y = 0 z = 1
x = 3 y = 1 z = 1
x = 3 y = 2 z = 1
x = 3 y = 3 z = 1
x = 3 y = 4 z = 1
x = 3 y = 5 z = -1.45682e+144
x = 4 y = 0 z = 1
x = 4 y = 1 z = 1
x = 4 y = 2 z = 1
x = 4 y = 3 z = 1
x = 4 y = 4 z = 1
x = 4 y = 5 z = -1.45682e+144

---

Do you Yahoo!?
The New Yahoo! Shopping - with improved product search

[Luigi Ballabio-2]

At 12:41 AM 10/14/03, amar singh wrote:

>I observed a strange thing in using the class.
>Am I using it wrongly, or could there possibly be some problem?
>
>Following is the code:
>     {
>       std::vector<double> x(5);
>      ...
>      std::vector<double> y(6);
>      ...
>      Math::Matrix z(x.size(),y.size());

Amar,
         we'll have to document it better. When I wrote it, I visualized
the thing as on two Carthesian axes:

   [yM] [                   ]
   |. | |                   |
   |. | |     Z matrix      |
   |. | |                   |
   |y2| |                   |
   [y1] [                   |
        [ x1  x2  . . .  xN ]

Therefore, the i-th y-value corresponds to the i-th row, and the j-th
x-value to the j-th column. The constructor of Matrix being:

     Matrix(Size rows, Size columns);

the constructor call in your code should be:

     Matrix(y.size(), x.size())

Please, if anyone feels it's the wrong way around, just tell me so that we
can change the parameter order...

Later,
         Luigi

[amar singh]

Ligui,

 

Thank you very much for your help, and sorry for creating the confusion.

It does work perfect, only I was not aware .

 

Thanks,

Amar

Luigi Ballabio <[hidden email]> wrote:

At 12:41 AM 10/14/03, amar singh wrote:

>I observed a strange thing in using the class.
>Am I using it wrongly, or could there possibly be some problem?
>
>Following is the code:
> {
> std::vector x(5);
> ...
> std::vector y(6);
> ...
> Math::Matrix z(x.size(),y.size());

Amar,
we'll have to document it better. When I wrote it, I visualized
the thing as on two Carthesian axes:

[yM] [ ]
|. | | |
|. | | Z matrix |
|. | | |
|y2| | |
[y1] [ |
[ x1 x2 . . . xN ]

Therefore, the i-th y-value corresponds to the i-th row, and the j-th
x-value to the j-th column. The constructor of Matrix being:

Matrix(Size rows, Size columns);

the constructor call in your code should be:

Matrix(y.size(), x.size())

Please, if anyone feels it's the wrong way around, just tell me so that we
can change the parameter order...

Later,
Luigi




-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
SourceForge.net hosts over 70,000 Open Source Projects.
See the people who have HELPED US provide better services:
Click here: http://sourceforge.net/supporters.php
_______________________________________________
Quantlib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

---

Do you Yahoo!?
The New Yahoo! Shopping - with improved product search

[amar singh]

Luigi,

 

I have a minor suggestion, which could prevent mistakes and hard to track bugs in using the matrix - It should throw an error for out of range access.

I think it would be helpful in avoiding mistakes in use.eg., If the matrix element accesses a pointer outside its allocation range, it could be quite upredictable and very hard to debug.

 

BTW, your 'state of the art'  Matrix class looks really excellent!

 

Regards,

Amar

amar singh <[hidden email]> wrote:

Ligui,

 

Thank you very much for your help, and sorry for creating the confusion.

It does work perfect, only I was not aware .

 

Thanks,

Amar

Luigi Ballabio <[hidden email]> wrote:

At 12:41 AM 10/14/03, amar singh wrote:

>I observed a strange thing in using the class.
>Am I using it wrongly, or could there possibly be some problem?
>
>Following is the code:
> {
> std::vector x(5);
> ...
> std::vector y(6);
> ...
> Math::Matrix z(x.size(),y.size());

Amar,
we'll have to document it better. When I wrote it, I visualized
the thing as on two Carthesian axes:

[yM] [ ]
|. | | |
|. | | Z matrix |
|. | | |
|y2| | |
[y1] [ |
[ x1 x2 . . . xN ]

Therefore, the i-th y-value corresponds to the i-th row, and the j-th
x-value to the j-th column. The constructor of Matrix being:

Matrix(Size rows, Size columns);

the constructor call in your code should be:

Matrix(y.size(), x.size())

Please, if anyone feels it's the wrong way around, just tell me so that we
can change the parameter order...

Later,
Luigi




-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
SourceForge.net hosts over 70,000 Open Source Projects.
See the people who have HELPED US provide better services:
Click here: http://sourceforge.net/supporters.php
_______________________________________________
Quantlib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

---

Do you Yahoo!?
The New Yahoo! Shopping - with improved product search

---

Do you Yahoo!?
The New Yahoo! Shopping - with improved product search